NCJ Number
217858
Journal
Journal of Experimental Criminology Volume: 2 Issue: 1 Dated: Spring 2006 Pages: 99-111
Date Published
2006
Length
13 pages
Annotation
This article reports on an electronic "handsearch" (the visual inspection of the contents of an article) of every available issue of the British Journal of Criminology (1960-2004), in order to determine how many randomized field experiments were published. These results were compared with earlier manual handsearch efforts to augment the Campbell Collaboration Social, Psychological, Educational, and Criminological Trials Register (C2-SPECTR).
Abstract
The electronic handsearch found only nine trials that involved randomized field experiments, although two others used quasi-random allocation such as alternation. Only one randomized field experiment had been published in the past 20 years, even though randomized trials is the design of choice for many who advocate evidence-based policy. The discrepancy between the number of randomized trials included in the C2-SPECTR and the current search (14 compared to 9) was due largely to inclusion criteria. C2-SPECTR handsearchers included a few trials in which items on questionnaires or vignettes were randomized to survey respondents. The authors speculate that the low number of randomized field experiments published in the British Journal of Criminology reflects the low production of such studies in the United Kingdom in general. This suggests a trend away from controlled research in criminal justice in favor of alternative and more naturalistic methods for studying interventions. Some researchers have argued that the randomized experiment has a number of characteristics that make it difficult to use in studying institutional treatment programs. 1 table, 8 notes, and 25 references